Analysis of vehicle motion under conditions of lateral wind disturbance is important for evaluating handling properties and vehicle stability. In the analysis, identification of vehicle dynamics is often carried out, and data for the identification is usually measured by a test with a lateral wind generator. However, vehicle transient response in the test usually converges for a short duration because of the limitation of the wind width. If the identification carried out from this data by conventional methods such as FFT, fine frequency resolution is not gained.In this research, an identification method based on the autoregressive model (AR-method), which is robust for a phenomenon of short duration, has been applied to the analysis of vehicle dynamics under the conditions in order to solve the above issue. As a result, fine frequency resolution of the transfer function has been obtained by the AR-method,and the time-series data estimated by the AR-method coincided with the experimental one.